sched/fair: remove util_est boosting
authorVincent Guittot <vincent.guittot@linaro.org>
Thu, 6 Jul 2023 13:51:44 +0000 (15:51 +0200)
committerPeter Zijlstra <peterz@infradead.org>
Wed, 26 Jul 2023 10:28:50 +0000 (12:28 +0200)
commitc2e164ac33f75e0acb93004960c73bd9166d3d35
tree88684a82dfd9821fc43e123a54f6a2505950a92a
parent752182b24bf4ffda1c5a8025515d53122d930bd8
sched/fair: remove util_est boosting

There is no need to use runnable_avg when estimating util_est and that
even generates wrong behavior because one includes blocked tasks whereas
the other one doesn't. This can lead to accounting twice the waking task p,
once with the blocked runnable_avg and another one when adding its
util_est.

cpu's runnable_avg is already used when computing util_avg which is then
compared with util_est.

In some situation, feec will not select prev_cpu but another one on the
same performance domain because of higher max_util

Fixes: 7d0583cf9ec7 ("sched/fair, cpufreq: Introduce 'runnable boosting'")
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lore.kernel.org/r/20230706135144.324311-1-vincent.guittot@linaro.org
kernel/sched/fair.c